CBL - Campus del Baix Llobregat

Projecte llegit

Títol: Automated design of planar structures with optimized radiation or scattering properties

Estudiants que han llegit aquest projecte:

Director/a: ÚBEDA FARRE, EDUARD

Departament: TSC

Títol: Automated design of planar structures with optimized radiation or scattering properties

Data inici oferta: 10-02-2022     Data finalització oferta: 10-10-2022


Estudis d'assignació del projecte:
    DG ENG AERO/SIS TEL
Tipus: Individual
 
Lloc de realització: EETAC
 
Paraules clau:
Genetic Algorithms; Discretization; Radar Cross Section
 
Descripció del contingut i pla d'activitats:
The computational design of planar perfectly conducting structures with improved Radar Cross Section signature has received special attention over the last decade. Thanks to these numerical tools, improved designs of car plates in the automotive industry or Radar antennas and radomes in the aviation industry have been obtained.

These schemes are normally based on the iterative and intensive search of improved designs computed with numerical scattering analysis tools. A brute-force search process of the improved design becomes time consuming and computationally expensive, whereby a global optimization tool that develops an intelligent search of the solution is normally implemented. In practice, at each iterative step of the design process, the new structure under analysis needs to be meshed and analyzed anew, which is also computationally demanding.

In this project, the Electric-Field Integral Equation (EFIE) is discretized to compute the scattering properties of perfectly conducting planar structures, meshed with triangular or quadrangular meshes. As global optimizing tool to lead the search of the solution the Genetic Algorithm strategy is proposed.

The conventional design-schemes requires to re-mesh the design at each step to ensure that adjacent facets of the mesh share one single edge (i.e. conformal mesh). Recently, it has been proposed an EFIE discretization that allows the accurate analysis of perfectly conducting structures through non-conformal meshes, where adjacent facets do not require a single edge in common. This is advantageous in practice because each mesh at each iterative step can be formed by combining small mesh-entities defined beforehand, with no need to impose conformality across the edges arising from the discretization. This project proposes to implement such novel strategy in the analysis of perfectly conducting planar structures.
 
Overview (resum en anglès):
The visibility that an object has with respect to a Radar is called the Radar Cross Section (RCS), and it is obtained from the electromagnetic dispersion in the far field when the object is struck by an electromagnetic plane wave. In order to be able to calculate the RCS of an arbitrary object, the surface of the object is meshed and basis functions of current discretization are applied, from which an approximation of the RCS is obtained. Typically, the basis functions adopted in triangular meshes are the RWG functions.

The design and synthesis of new structures with enhanced electromagnetic signature requires on the one hand the meshing and calculation of the RCS for many possible candidates, and on the other, the choice of an itinerary of intermediate solutions that leads to an almost optimal solution. In practice, this procedure is computationally expensive.

Recently, a procedure has been developed that streamlines the analysis of the electromagnetic scattering of planar conductive structures. In this work, this tool has been used to efficiently search for a solution with an improved RCS. This tool, which we call the thick-surface approach, makes it possible to obtain the mesh of any complex planar structure from the concatenation and displacement of the known mesh of a fundamental basic cell. This is an advantage over traditional methods, which required re-meshing for any composite planar structure.

The objective of this work is to find the optimal distribution that the conducting planar structure
should have so that its RCS becomes maximum. For this purpose, a matrix is used to define where we want the initial cell to be cloned. Using genetic algorithms, an example of a stochastic optimization algorithm, the most suitable distribution of this position matrix to have the maximum RCS is searched for and found.

In this work, the optimization is carried out through the Matlab program, where the necessary code is written to be able to carry out the simulations.

© CBLTIC Campus del Baix Llobregat - UPC